Photoelectric conversion devices employing CIGS semiconductors are conventionally known. For example, PTL1 discloses a solar cell which includes an electrode of a molybdenum (Mo) layer formed by sputtering, a photoelectric conversion layer or a light absorbing layer of a Cu(In1-xGax)Se2 film, a buffer layer of a CdS layer formed by a solution growth method, a window layer of ZnO deposited by an MOCVD method, and an electrode of aluminum formed by vapor deposition or sputtering and patterning, these layers being provided in this order on a substrate of soda lime glass (SLG).
The photoelectric conversion layer is formed by a so-called three-step method, which includes the steps of: depositing a film of In—Ga—Se on the substrate; supplying Cu—Se to the In—Ga—Se film with heating to convert the In—Ga—Se film into a film of Cu—Se-rich Cu(In1-xGax)Se2 containing excess Cu—Se; and supplying In—Ga—Se to the Cu—Se-rich Cu(In1-xGax)Se2 film to convert the excess Cu—Se into Cu(In1-xGax)Se2 to form a Cu(In1-xGax)Se2 film (wherein 0<x≦1) and then removing a Cu—Se compound from a surface of the Cu(In1-xGax)Se2 film.